Test Bank Chapter 4 Forecasting With Regression Trends

Forecasting and Predictive Analytics with Forecast X, 7e (Keating)

Chapter 4 Extrapolation 2. Introduction to Forecasting with Regression Trend Models

1) The least squares procedure minimizes the sum of

A) the residuals.

B) squared maximum error.

C) absolute errors.

D) squared residuals.

E) None of the options are correct.

2) A residual is

A) the difference between the mean of Y conditional on X and the unconditional mean.

B) the difference between the mean of Y and its actual value.

C) the difference between the regression prediction of Y and its actual value.

D) the difference between the sum of squared errors before and after X is used to predict Y.

E) None of the options are correct.

3) The condition expectation of a random variable

A) is denoted by E(Y | X = x) and tells us the expected value of Y given a particular value of X.

B) is the foundation of the simple regression model.

C) is modeled using a linear function in the simple regression model.

D) is represented graphically by the regression line in the simple bivariate model.

E) All of the options are correct.

4) The Y-intercept of the simple regression model

A) rarely has a useful interpretation.

B) almost always has a useful interpretation.

C) is always a positive number.

D) is always positive when the correlation between the dependent and independent variable is positive.

E) All of the options are correct.

5) The Y-intercept of a regression line is −14, and the slope is 3.5. Which of the following is not correct?

A) When Y increases by one, X increases by 3.5.

B) When X increases by one, Y increases by 3.5.

C) The regression line crosses the Y-axis at −14.

D) X and Y are positively related.

E) None of the options are correct.

6) Income is used to predict savings. For the regression equation Y = 1,000 + .10X, which of the following is true?

A) Y is income, X is savings, and income is the independent variable.

B) Y is income, X is savings, and savings is the independent variable.

C) Y is savings, X is income, and savings is the independent variable.

D) Y is savings, X is income, and income is the independent variable.

7) The sign on the slope estimate in a regression problem

A) is the same as the sign of the Y-intercept.

B) is the opposite of the sign of the correlation of Y and X.

C) has no relationship to the sign of the correlation of Y and X.

D) always has the same sign as the correlation of Y and X.

E) None of the options are correct.

8) X-Y data have been collected in which X ranges between 50 and 100 and Y ranges between 1200 and 2000. It is not wise to use the resulting regression line equation to predict Y when X is equal to −10 because

A) a negative number cannot be used.

B) the predicted value for Y might turn out to be negative.

C) the Y-intercept might be above zero.

D) the proposed X value is well beyond the range of observed data.

E) All of the options are correct.

9) The following regression equation was estimated: Y = −2.0 + 4.6X. This indicates that

A) there has been an error since "b" cannot be a negative number.

B) there is a negative relationship between the two variables.

C) Y equals 44 when X is 10.

D) the correlation coefficient for Y and X will be negative.

E) All of the options are correct.

10) Which of the following is not a reason to employ simple linear regression to generate sales forecasts for a retail outlet store?

A) Causal relationships can be examined.

B) Trend can be handled using a time index.

C) Data seasonality can be handled by deseasonalizing the data.

D) The conditional mean of sales can be estimated.

E) None of the options are correct.

11) Sample regression model forecast errors are called

A) disturbances.

B) residuals.

C) least-squares predictions.

D) outliers.

E) None of the options are correct.

12) The regression slope term (β) in the simple bivariate regression model is

A) correctly interpreted as dY/dX.

B) usually known to the investigator.

C) the change in the conditional mean of Y given a unit change in X.

D) undetermined using the OLS method.

E) None of the options are correct.

13) Regression model disturbances (forecast errors)

A) are assumed to follow a normal probability distribution.

B) are assumed to be independent over time.

C) are assumed to average to zero.

D) can be estimated by OLS residuals.

E) All of the options are correct.

14) Serial correlation (or autocorrelation) causes estimates of the

A) slope parameter β to be understated on average.

B) slope parameter β to be overstated on average.

C) estimates of the standard errors to be understated on average.

D) estimates of the standard errors to be overstated on average.

E) None of the options are correct.

15) Visual inspection of the data will help the forecaster identify

A) trend.

B) seasonality.

C) linearity.

D) nonlinearity.

E) All of the options are correct.

16) Which of the following is a tool used in model selection?

A) Seasonality

B) Cyclicity

C) Growth

D) Plotting the data

E) None of the options are correct.

17) Which of the following is not a recommended step in preparing a forecast using the simple linear regression model?

A) Visually inspect the data.

B) Forecast the independent variable.

C) Specify a regression model.

D) Select a holdout period for model evaluation.

E) None of the options are correct.

18) Fit and accuracy

A) are the same things.

B) do not depend on the fitted regression model.

C) do not depend on the estimated standard error.

D) reflect in sample versus out-of-sample model forecast errors.

E) All of the options are correct.

19) Which of the following is not a method for estimating data with trend?

A) Holt's smoothing

B) Winter's smoothing

C) Regression linear trend model

D) Using time as the independent variable

E) None of the options are correct.

20) The most common mathematical trend equation for a time series is called the least squares trend because it is the line which minimizes the sum of the

A) squares of deviations from the sample mean.

B) deviations from the mean.

C) squared vertical deviations from the trend line.

D) deviations from the mean of the X variable.

E) None of the options are correct.

21) Consider the following model: Sales = α + β(TIME)2 + ε. The trend is modeled here as a(n)

A) linear trend.

B) exponential trend.

C) quadratic trend.

D) logarithmic trend.

E) None of the options are correct.

22) Seasonal indices of sales for the Black Lab Ski Resort are for 1.20 for January and .80 for December. If December sales for 1998 were $5,000, a reasonable estimate of sales for January 1999 is

A) $4,800.

B) $6,000.

C) $7,500.

D) $10,000.

E) None of the options are correct.

23) The expected trend value of September sales for a firm is $900. Assuming a September seasonal index of .91, what would be the seasonally adjusted forecast for September?

A) $989

B) $950

C) $900

D) $819

E) None of the options are correct.

24) Which of the following is not correct about causal regression analysis of the form Y = f(X)?

A) Selection of the appropriate causal variable Y is important.

B) Selection of the appropriate causal variable X is important.

C) Use of past experience to identify X is common.

D) Use of economic theory to identify X is common.

E) All of the options are correct.

25) Consider the following model linking seasonally adjusted retail-store sales (RSSA) to disposable personal income (DPI):

RSSA = −1,813,520 + 127.429(DPI).

If the quarter four seasonal index is 1.07264 and DPI is 19,119.6, our forecast for quarter-four sales is

A) $550,200.51.

B) $619,352.99.

C) $620,531.65.

D) $668,116.89.

E) None of the options are correct.

26) Big Daddy

Total annual sales for BDC in 2001 are forecasted at $120 million. Based on the seasonal indexes above, sales in the first three months of 2001 should be

A) 10 million dollars.

B) 1.2 million dollars.

C) 30 million dollars.

D) 31 million dollars.

E) None of the options are correct.

27) Big Daddy

If December 2000 sales for BDC are 20 million, what is a reasonable estimate for sales in January of 2001?

A) 16 million dollars

B) 19.2 million dollars

C) 20.84 million dollars

D) 30 million dollars

E) None of the options are correct.

28) Big Daddy

If BDC sales in November of 2000 were 12 million dollars, November sales after adjustment for seasonal variation are

A) 10.91 million dollars.

B) 13.2 million dollars.

C) 13.1 million dollars.

D) Not enough information is present to answer the question.

E) None of the options are correct.

29) Which of the following statements is true?

A) Autocorrelation arises when there is a perfect linear association among the independent variables in the sample.

B) Autocorrelation and its presence have no effect on the Gauss-Markov theorem.

C) Autocorrelation causes the sum of squares decomposition to become unreliable.

D) Autocorrelation causes the ordinary least squares estimate of the error variance to become biased.

E) None of the options are correct.

30) Which of the following is incorrect?

A) R2 is a measure of the degree of variability in the dependent variable about its sample mean explained by the regression line.

B) R2 measures the "goodness-of-fit" of a regression model.

C) The null hypothesis that R2 = 0 can be tested using an F-test.

D) The best model selection criteria and variable selection criteria for a forecaster to use is the maximization of R Squared.

E) None of the options are correct.

31) The autocorrelation parameter is used to measure

A) disturbances or independent random variates.

B) correlation between residuals.

C) the slope of the regression line.

D) error or difference between a data point and the regression line.

E) difference between present and past residuals.

32) Which of the following is not used to solve the problem of autocorrelation?

A) Autoregressive models

B) Improving the model specification

C) Moving average smoothing

D) First differencing the data

E) Regression using percentage changes

33) If the residuals in a regression equation are positively autocorrelated, which of the following is not a problem when the least squares procedure is used?

A) The standard error of the regression slope coefficient underestimates the true variability of the estimated regression.

B) Confidence intervals are no longer strictly applicable.

C) The t and F distributions are no longer strictly applicable.

D) The regression coefficients are no longer strictly applicable.

E) The standard error of the regression seriously understates the variability of the error terms.

34) When autocorrelation is present, which of the following is not a problem?

A) The F-statistic may be unreliable.

B) The t-statistics for each coefficient will be overstated.

C) The R-squared statistic may be unreliable.

D) The estimated standard errors will be larger than the true standard errors.

E) None of the options are correct.

35) When severe autocorrelation is indicated after a regression model has been estimated, which underlying regression assumption is violated?

A) The population of Y values is normally distributed about the population regression line.

B) The dispersion of population data points around the population regression line remains constant everywhere along the line.

C) The error terms are independent of each other.

D) A linear relationship exists between X and Y in the population.

E) Heteroscedasticity

36) One method for solving the autocorrelation problem is to take advantage of the correlation between adjacent observations. This method is called

A) Regression on percentage changes in the data.

B) Multiple regression.

C) Durbin-Watson model.

D) Cochrane-Orcutt regression method.

E) Exponential model.

37) Which of the following is not an indicator of regression fit?

A) Does the estimated sign of the slope coefficient make economic sense?

B) Is R-squared greater than one?

C) Is the model underspecified?

D) Are X and Y significantly related?

E) All of the options are correct.

38) Testing the null hypothesis that the slope coefficient is zero uses what sampling distribution for small sample sizes?

A) Normal.

B) Chi-square.

C) t distribution with n-1 degrees of freedom.

D) Standard Normal.

E) None of the options are correct.

39) Which diagnostic test allows the researcher to claim that her model explains x-percent of the variation in the dependent variable?

A) Durbin-Watson test

B) Coefficient of Determination

C) t-test on slope coefficient

D) Sum of squared residuals

E) None of the options are correct.

40) Which of the following would indicate a perfect model fit?

A) R2 = 1

B) R2 = 0

C) Durbin-Watson = 2

D) t-test for slope > 2

E) None of the options are correct.

41) Consider the following time trend regression model for explaining the behavior of disposable personal income (DPI): DPI = 17,000 + 41(TIME). If the regression standard error were 150, what is an approximate 95% prediction interval for quarter 3 DPI?

A) 16,366 to 17,105

B) 16,823 to 17,423

C) 16,932 to 18,108

D) 17,102 to 18,345

E) None of the options are correct.

42) Serial correlation violates which classical assumption?

A) The error terms have a zero mean.

B) The error terms follow a normal distribution.

C) The error terms are independent of each other.

D) The error terms have the same variance.

43) Which of the following does not become unreliable when serial correlation is present?

A) R-squared

B) t-tests

C) OLS slope estimates

D) Error sum of squares

E) None of the options are correct.

44) Autocorrelation in a regression model occurs when there is some correlation

A) among the explanatory variables.

B) among the residuals and the values of the explanatory variables.

C) between the error in one period and the error in the next period.

D) between the slope and intercept estimates.

E) None of the options are correct.

45) If R2 is .95 in a simple regression model, it can be said that

A) X and Y have a correlation of .95.

B) the relationship between X and Y is positive.

C) 5 percent of Y's variability is caused by variability in X.

D) 95% of Y's variability can be explained by X's variability.

46) The autocorrelation parameter defined as

is used to measure

A) disturbances of independent random variables.

B) correlation between regression error terms.

C) the Durbin-Watson statistic.

D) the difference between the forecast and the estimated regression line.

E) None of the options are correct.

47) Which of the following is not true regarding testing for serial correlation?

A) The DW statistic has a range from 0 to 1.

B) The DW statistic has a range from 0 to 4.

C) If positive serial correlation exists, the value of DW approaches zero.

D) If negative serial correlation exists, the value of DW approaches 4.

E) None of the options are correct.

48) Autocorrelation leads to or causes

A) heteroscedasticity.

B) serial correlation.

C) spurious regression.

D) nonlinear regression.

E) All of the options are correct.

49) The Durbin-Watson statistic is based upon the

A) estimated residuals.

B) estimated Y-intercept term.

C) estimated slope coefficient.

D) fitted regression line.

E) All of the options are correct.

50) In applying the Durbin-Watson test, the required inputs are

A) DW lower bound.

B) sample size.

C) number of independent variables.

D) estimated DW value.

E) All of the options are correct.

51) What possible decisions can be made using the Durbin-Watson test?

A) Reject serial correlation.

B) Conclude positive serial correlation.

C) Conclude negative serial correlation.

D) Conclude nothing.

E) All of the options are correct.

52) How many critical regions does the Durbin-Watson test have?

A) one

B) two

C) three

D) four

E) five

53) Which of the following is not a possible cause of serial correlation?

A) Trends in the data

B) Data seasonality

C) Periodic cycles in the data

D) Leaving out an important explanatory variable

E) None of the options are correct. (i.e. they could all be possible causes).

54) Which of the following is not a possible cure for serial correlation?

A) Detrend the data by first differencing.

B) Deseasonalize the data.

C) Introduce a lagged value of the dependent variable.

D) Introduce a squared value of the independent variable.

E) None of the options are correct.

55) First-differencing the data is a way to

A) detrend the data.

B) reseasonalize the data.

C) remove heteroscedasticity from the data.

D) remove any data nonlinearities.

56) What is a common way forecasters attempt to eliminate heteroscedasticity?

A) Transform the data by first differencing.

B) Deseasonalize the data.

C) Transform the data using logarithms.

D) Apply Ordinary Least Squares.

E) All of the options are correct.

57) What pattern of residuals in a time-series plot would indicate heteroscedasticity?

A) Purely random

B) Elliptical

C) Circular

D) Funnel shaped

E) Linear

58) Stock price data show periods of relatively calm interrupted by periods of enhanced price volatility. This suggests stock price data are

A) homoscedastic.

B) autocorrelated.

C) nonlinear.

D) heteroscedastic.

E) linear.

59) Which of the following is a problem caused by heteroscedasticity when using ordinary least squares?

A) The Y-intercept estimate is biased.

B) The slope estimate is biased.

C) The regression standard error estimate is biased.

D) The slope estimate is unbiased.

60) Exact prediction intervals for the dependent variable

A) are bow-shaped around the estimated regression line.

B) are linear around the estimated regression line.

C) do not take the variability of Y around the sample regression into account.

D) do not take the randomness of the sample into account.

E) None of the options are correct.

61) What is the approximate 95% prediction interval for the dependent variable when the independent variable value is 20, assuming the fitted regression line is: Y = 1.50 + 6.0(X). Assume the sample size is 20 and the standard error of the regression (SEE) is 1.2. You should use the "rule of thumb" used in class here.

A) ~117.84 to 126.16

B) ~119.32 to 124.12

C) ~108.60 to 137.40

D) ~17.65 to 22.35

E) ~119.10 to 123.90

62) In a regression problem, if R-squared is equal to .90, it means that the ratio of the sum of

A) squared residuals to the sum of squared Y values around the mean of Y is .90.

B) squared Y values around the mean of Y to the sum of squared residuals is .90.

C) squared residuals to the sum of squared Y values around the mean of Y is .10.

D) squared Y values around the mean of Y to the sum of squared residuals is .10.

E) None of the options are correct.

63) In a cross-sectional study of sales in different cities, the following relationship between sales revenue (S = sales revenue in dollars) and city size as measured by population (POP = population in thousands) was estimated: S = 37.02 + 0.6734(POP). What is the approximate 95 percent prediction interval for sales in a city of 155,000 people when the regression standard error is 32.7214? (Note: you should use a rule of thumb here)

A) 46.721 to 254.586.

B) 75.955 to 206.839.

C) 95.899 to 187.333.

D) 108.676 to 174.118.

64) What is not likely to be a problem when applying ordinary least squares to cross-sectional data?

A) Nonlinearity.

B) Unbiased coefficient estimates.

C) Autocorrelation.

D) Heteroscedasticity.

E) None of the options are correct.

65) Cross-sectional regression models linking personal disposable income to consumption expenditure are likely to be hampered by

A) nonlinearity.

B) homoscedasticity.

C) autocorrelation.

D) heteroscedasticity.

E) None of the options are correct.

66) The following regression represents seasonally adjusted shoe store sales in millions of dollars in the United States.

Audit Trail — Coefficient Table (Multiple Regression Selected

Consider the simple regression above. The dependent variable is seasonally adjusted shoe store sales in the United States. The independent variable is the index of time.

Does it appear that there is a trend to shoe store sales?

A) Yes, the slope term is positive and statistically significant.

B) No, although the slope term is positive, it is statistically insignificant.

C) There is no way to tell from the data given whether there is a trend to shoe store sales.

D) None of the options are true.

67) The following regression represents seasonally adjusted shoe store sales in millions of dollars in the United States.

Audit Trail — Coefficient Table (Multiple Regression Selected

Consider the simple regression above. The dependent variable is seasonally adjusted shoe store sales in the United States. The independent variable is the index of time.

In the shoe store sales regression above, the reported Durbin-Watson statistic

A) indicates that there is no serial correlation.

B) indicates that the seasonal adjustment may not have eliminated all the seasonality.

C) indicates that the seasonal adjustment did eliminate any seasonal effect.

D) proves that trend does not exist in the shoe store sales data.

68) The following regression represents seasonally adjusted shoe store sales in millions of dollars in the United States.

Audit Trail — Coefficient Table (Multiple Regression Selected

Consider the simple regression above. The dependent variable is seasonally adjusted shoe store sales in the United States. The independent variable is the index of time

The coefficient of determination in the shoe store sales regression

A) indicates that the independent variable is statistically significant.

B) indicates that there is no serial correlation in the error terms.

C) indicates that the confidence interval for this regression would be plus and minus two times 0.12.

D) indicates that about 77 percent of the variation in shoe sales is explained.

69) The following regression represents seasonally adjusted shoe store sales in millions of dollars in the United States.

Audit Trail — Coefficient Table (Multiple Regression Selected

Consider the simple regression above. The dependent variable is seasonally adjusted shoe store sales in the United States. The independent variable is the index of time.

The Durbin Watson (12) statistic is reported because

A) the data has 12 observations available for analysis.

B) there are 12 degrees of freedom available for calculating the significance of the t-statistic.

C) the time series was probably monthly data.

D) the researcher is examining multicollinearity.

70) The following regression represents seasonally adjusted shoe store sales in millions of dollars in the United States.

Audit Trail — Coefficient Table (Multiple Regression Selected

Consider the simple regression above. The dependent variable is seasonally adjusted shoe store sales in the United States. The independent variable is the index of time.

The reported p-value for the time index

A) indicates that the time variable is statistically insignificant.

B) indicates that the time variable is statistically significant.

C) indicates that the time variable is unrelated to shoe sales.

D) indicates that there is little explanatory power to this regression model.

71) The following regression represents seasonally adjusted shoe store sales in millions of dollars in the United States.

Audit Trail — Coefficient Table (Multiple Regression Selected

Consider the simple regression above. The dependent variable is seasonally adjusted shoe store sales in the United States. The independent variable is the index of time.

The slope term in this regression

A) is 1523.74.

B) is 2.90.

C) is 0.13.

D) is 11.19.

72) The following regression represents seasonally adjusted shoe store sales in millions of dollars in the United States.

Audit Trail — Coefficient Table (Multiple Regression Selected

Consider the simple regression above. The dependent variable is seasonally adjusted shoe store sales in the United States. The independent variable is the index of time.

The data in this regression covers the period 1992 through 2004. Does the regression seem to pass the "first quick check"?

A) No, there are too few observations to make a judgement.

B) No, there is no way to perform this test without extra information.

C) Yes, there is ample information to make this judgement.

D) Yes, both t-statistics are greater than two.

73) - The following output resulted from a regression model where SAGap is seasonally adjusted Gap sales and dpi is disposable income per capita.

Audit Trail — Coefficient Table (Multiple Regression Selected

A) This regression model is a causal model.

B) This regression model is a nonlinear model.

C) This regression model is a multiple regression model.

D) This regression model is a lagged model.

E) None of the options are true.

74) The following output resulted from a regression model where SAGap is seasonally adjusted Gap sales and dpi is disposable income per capita.

Audit Trail — Coefficient Table (Multiple Regression Selected

Linear least squares regression chooses values for the intercept and slope to minimize

A) the sum of the squared errors (distances) from the reference line.

B) the sum of the squared errors (distances) from the regression line.

C) the sum of the squared errors (distances) from the average value of the dependent variable.

D) the R2 (coefficient of determination) of the regression.

E) None of the options are correct.

75) Confidence Bands

Consider the diagram above representing the confidence bands around a linear least squares regression line.

A) The lines AB and CD represent the true 95% confidence bands.

B) The lines EF and GH represent the true 95% confidence bands.

C) The shaded area represents the true 95% confidence band.

D) Neither set of lines nor the shaded area represent the true 95% confidence band.

76) Confidence Bands

Use the confidence band diagram above.

A) An approximate 95% confidence interval for a regression estimate would appear like lines AB and CD.

B) An approximate 95% confidence interval for a regression estimate would appear like lines EF and GH.

C) Neither set of lines (AB and CD nor EF and GH) represents a reasonable depiction of a confidence interval.

D) Confidence intervals for least squares linear regression estimates may take on many shapes that appear quite unlike those depicted in the diagram.

77) Confidence Bands

Given that we have collected pairs of observations on two variables X and Y, we would consider fitting a straight line with X as an explanatory variable if:

A) the change in Y is an additive constant.

B) the change in Y is a constant for each unit change in X.

C) the change in Y is a fixed percent of Y.

D) the change in Y is exponential.

E) None of the options are correct.

78) The least squares regression line is the line:

A) which is determined by use of a function of the distance between the observed Y 's and the predicted Y's.

B) which has the smallest sum of the squared residuals of any line through the data values.

C) for which the sum of the residuals about the line is zero.

D) which has all of the above properties.

E) which has none of the above properties.

79) For children, there is approximately a linear relationship between "height" and "age". One child was measured monthly. Her height was 75 cm at 3 years of age and 85 cm when she was measured 18 months later. A least squares line was fit to her data. The slope of this line is approximately:

A) 0.55 cm/m.

B) 10 cm/m.

C) 25 cm/m.

D) 1.57 cm/m.

E) 2.1 cm/m.

80) There is an approximate linear relationship between the height of females and their age (from 5 to 18 years) described by:

height = 50.3 + 6.01(age)

where height is measured in cm and age in years. Which of the following is not correct?

A) The estimated slope is 6.01 which implies that children increase by about 6 cm for each year they grow older.

B) The estimated height of a child who is 10 years old is about 110 cm.

C) The estimated intercept is 50.3 cm which implies that children reach this height when they are 50.3/6.01=8.4 years old.

D) The average height of children when they are 5 years old is about 50% of the average height when they are 18 years old.

E) My niece is about 8 years old and is about 115 cm tall. She is taller than average.

81) A study was conducted to examine the quality of fish after seven days in ice storage. For this study

Y = measurement of fish quality (on a 10-point scale with 10 = BEST.)

X = # of hours after being caught that the fish were packed in ice.

The sample linear regression line is: Y = 8.5 - .5X. From this we can say that:

A) a one-hour delay in packing the fish in ice decreases the estimated quality by .5.

B) a one-hour delay in packing the fish in ice increases the estimated quality by .5.

C) if the estimated quality increases by 1, then the fish have been packed in ice one hour sooner.

D) if the estimated quality increases by 1, then the fish have been packed in ice two hours later.

E) Can't really say until we see a plot of the data.

82) The yield of a grain, Y (t/ha), appears to be linearly related to the amount of fertilizer applied, X (kg/ha). An experiment was conducted by applying different amounts of fertilizer (0 to 10 kg/ha) to plots of land and measuring the resulting yields. The following estimated regression line was obtained:

yield = 4.85 + .05(fertilizer)

Which of the following is not correct?

A) If no fertilizer was used, the yield is estimated to be 4.85 t/ha (note that zero use of fertilizer is within the range of this data).

B) If fertilizer is applied at 10 kg/ha, the estimated yield is 5.35 t/ha.

C) For every additional kg/ha of fertilizer applied, the yield is estimated to increase 0.05 t/ha.

D) To obtain an estimated yield of 5.2 t/ha, you need to apply 7.0 kg/ha of fertilizer.

E) If the current level of fertilizer is changed from 7.0 to 9.0 kg/ha, the yield is estimated to increase by 0.20 t/ha.

83) In cross-sectional regression analysis,

A) the time index is the only independent variable used.

B) serial correlation is a serious problem.

C) only a single independent variable may be used.

D) the data all pertain to one time period.

84) The Durbin-Watson statistic

A) takes on values from 0 to 4.

B) tests for serial probability.

C) tests the coefficient(s) of the independent variable(s).

D) cannot be used with fewer than 10 observations.

85) The Standard Error of the Estimate (SEE) in an ordinary least squares regression

A) is the standard error of the dependent variable.

B) is the standard error of the independent variable.

C) is used to test for autocorrelation.

D) is used to test for serial correlation.

86) The Y-intercept (b0) in an ordinary least squares regression represents the

A) predicted value of Y when X = 0.

B) change in the estimated average Y per unit change in X.

C) predicted value of Y.

D) variation around the sample regression line.

87) The standard error of the estimate in an ordinary least squares regression is a measure of the

A) total variation of the Y variable.

B) variation around the sample regression line.

C) explained variation.

D) variation of the X variable.

88) The coefficient of determination in an ordinary least squares regression tells us

A) that the correlation coefficient is larger than 1.

B) whether the correlation coefficient has any significance.

C) that we should not partition the total variation.

D) the proportion of the total variation that is explained.

89) If the plot of the residuals is fan shaped, which assumption of ordinary least squares regression is violated?

A) Normality

B) Homoscedasticity

C) Independence of errors

D) No assumptions are violated; the graph should resemble a fan.

90) If the correlation coefficient in an ordinary least squares regression = 1.00, then

A) all the data points must fall exactly on a straight line with a slope that equals 1.00.

B) all the data points must fall exactly on a straight line with a negative slope.

C) all the data points must fall exactly on a straight line with a positive slope.

D) all the data points must fall exactly on a horizontal straight line with a zero slope.

91) The strength of the linear relationship between two numerical variables may be measured by the

A) scatter diagram.

B) coefficient of determination.

C) slope.

D) Y-intercept.

92) In a linear least squares regression, what is one of the two parameters estimated?

A) the coefficient of determination

B) the dependent variable

C) the slope

D) the independent variable

93) In linear least squares regression, we use "squares" to overcome what difficulty?

A) because standard deviations will not be "standardized" unless squares are used.

B) because we often use time series data, using squares is required.

C) using squares reduces the size of the error terms in a regression analysis.

D) using squares counts both positive and negative errors rather than allowing them to offset one another.

94) Because every software package that includes regression analysis uses the "Normal Equations,"

A) all of the diagnostic statistics produced by the various packages will vary.

B) every software package will provide slightly different results.

C) every software package will provide identical results.

D) every package will include "Normalized Coefficients."

95)

In table 4.1 from the text, four different data sets are displayed along with the regressions associated with each data set. What point was being made in the text?

A) All the data sets are more alike than they are different.

B) While there are four different data sets, all differ only by some random variation.

C) While each data set is quite different from the rest, all result in the same regression.

D) Regression works equally well with almost any data set.

96) Table 4.2

With the data in Table 4.2 from the text, a time trend regression was run.

A) This regression used the time index as a dependent variable.

B) This regression used the time index as an independent variable.

C) This regression could be used to prove causality.

D) This regression was used as an example of "what not to do."

97)

This is a linear trend regression. Which of the following is not true?

A) There is a negatively sloped linear trend.

B) There is some explanatory power since the MAPE is only 1.24%.

C) There is a near vertical linear trend.

D) No improvement is shown from that of a naïve model since the Theil's U is just 1.89.

98)

The jewelry regression shown

A) is a linear time trend regression.

B) is a causal regression.

C) is a nonlinear regression.

D) is a seasonal regression.

99)

When we use the t distribution to evaluate the statistical significance of a regression coefficient,

A) we are assuming that the distribution of the coefficient is binomial.

B) we are calculating the R2.

C) we want zero to be at the center of the distribution.

D) we assume the distribution is centered on the calculated coefficient.